7. Modeling Evolution: Evolutionary Computation“How does evolution produce increasingly fit organisms in environments which are highly uncertain forindividual organisms? How does an organism use its experience to modify its behavior in beneficial ways (i.e.how does it learn or ‘adapt under sensory guidance’)? How can computers be programmed so that problem-solving capabilities are built up by specifying ‘what is to be done’ rather than ‘how to do it’?” [Holland, 1975,page 1]

These were some of the questions concerning John Holland when he thought of Genetic Algorithms (GA’s)in the 1960's. All these questions were shown to be reducible to a problem of optimizing multi-parameterfunctions. Nature’s “problem” is to create organisms that reproduce more (are more fit) in a particularenvironment: the environment-organism coupling dictates the selective pressures, and the solutions to thesepressures are organisms themselves. In the language of optimization, the solutions to a particular problem(say, an engineering problem), will be selected according to how well they solve that problem. GA’s areinspired by natural selection as the solutions to our problem are not algebraically calculated, but rather foundby a population of solution alternatives which is altered in each time step of the algorithm in order to increasethe probability of having better solutions in the population. In other words, GA’s, or other EvolutionaryStrategies (ES) such as Evolutionary Programming (EP), explore the multi-parameter space of solutionalternatives for a particular problem, by means of a population of encoded strings (standing for alternatives)which undergo variation (crossover and mutation) and are reproduced in a way as to lead the population toever more promising regions of this search space (selection) [Goldeberg, 1989; Mitchell, 1999; De Jong,2006].

7.1 Evolutionary Strategies and Self-Organization

The underlying idea of computational ES is the separation of solutions for a particular problem (e.g. amachine) from descriptions of those solutions (memory). GA’s work on these descriptions and not on thesolutions themselves, that is, variation is applied to descriptions, while the respective solutions are evaluated,and the whole (description-solution) selected according to this evaluation. Such machine/descriptionseparation follows von Neumann’s self-reproducing scheme (see chapter 6) which is able to increase thecomplexity of the (organization of) machines described. Therefore, the form of organization evolved by GA’sis not self-organizing in the sense of a boolean network or cellular automata (see chapter 4). Even though thesolutions are obtained from the interaction of a population of elements, and in this sense following the generalrules usually observed by computationally emergent systems (e.g. Langton [1988], Mitchell [1992]), theydo not self-organize since they rely on the selective pressures of some environment (in ES, defined by anexplicit or implicit fitness function). The order so attained is not a result of the internal dynamics of acollection of interacting elements, but is instead dictated by the external selection criteria. In this sense, ESfollow an organizing scheme that is driven by external selection of encoded symbolic descriptions (a “Turingtape” ). It is perhaps useful to think that ES are modeling the most fundamental design principle of biologicalsystems: natural selection. While self-organizing systems model the dynamical characteristics of matter, ESmodel the existence of, external, selective pressures on populations of symbolic descriptions of some system.While self-organization models material dynamics, ES models the selection of information about dynamics.

7.2 Development and morphogenesis: self-organization and selection come together

Since the original introduction of GA’s, many subsequent developments had to do with the inclusion of adevelopmental stage, or intermediate layers between genotype and phenotype; in other words, the creationof some artificial morphogenesis or regulation. The idea has been to encode rules that will themselves self-

28

organize to produce a phenotype, rather than the direct encoding of the phenotype itself, or the introductionof gene regulation . As discussed in class, these rules often use L-System grammars which dictate productionsystem programs [Wilson, 1988] leading to some phenotype. The most important advantage of thisintermediate stage, as explored by Kitano [1990], Gruau [1993], Belew [1992] and others, is the ability tocode for much larger structures than a direct encoding allows. In practical terms, they have solved some ofthe scalability problems of encoding (e.g.) neural networks in GA’s, by reducing the search spacedramatically.

L-system grammars are higher-level descriptions of self-organizing developmental processes. However, thesefirst approaches used solely context-free, state-determined, L-System grammars, compromising epistasis (ormutual, non-linear, influence of genetic descriptions amongst each other) in the simulation of self-organizingdevelopment. Dellaert and Beer [1994] and Kitano [1994], for instance, used Boolean networks to simulategenetic epistasis and self-organization. In other words, the GA encodes rules which construct Booleannetworks whose nodes stand for aspects of the phenotypes we wish to evolve on some physical simulation.In Dellaert and Beer’s model, the nodes stand for cell mitosis and other characteristics. This way, thesolutions of the GA are self-organizing systems whose attractor behavior dictates pre-defined phenotypictraits.

These approaches in effect offer an emergent morphology, that is, they encode rules which will themselvesself-organize into some phenotype (instead of strict programming of morphology). The indirect encodingfurther allows the search to occur in a reduced space, amplified through development. An interesting sideeffect has to do with the appearance of modularity traits on the evolved phenotypes [Wagner, 1995].Subsequent developments paid even more attention to the contextual regulation that indirect encodings affordto the search [Rocha 1995, 1997]. More recently, given our expanded view of genomics, other intermediatelayers between genotype and phenotype have been explored, such as transcription regulation [Reil, 1999;Hallinan & Wiles, 2004] and RNA Editing [Rocha et al, 2006]. The inclusion of more sophisticated regulationof genetic information prior to translation, while not necessarily including a self-organizing component,allows us to model a much more realistic genotype/phenotype/environment interaction. Instead of genotypesused exclusively for Mendelian inheritance (see chapter 5) of (directly encoded) phenotypic traits, ES withgenotype regulation allow us to model the contextual, plastic development of phenotypes we have come tounderstand via modern Genomics—thus also learning additional design principles for bio-inspiredcomputation [Huang et al, 2007].

The most important aspect of GA’s with emergent morphologies is the utilization in the same model of anexternal selection engine (the GA) coupled to a particular self-organizing dynamics (e.g. Boolean networks)standing for some materiality. Such schemes bring together, computationally, the two most important aspectsof evolutionary systems: self-organization and selection. These models belong to a category of self-organization referred to as Selected Self-Organization which is based on symbolic memory [Rocha, 1996,1997, 1998]. Selected Self-Organization with distributed memory is also possible in autocatalytic structures,though its evolutionary potential is much smaller than the local memory kind [Rocha, 2001][Vasas, 2010].The reason lies in Von Neumann’s notion of self-reproduction (see chapter 6). The introduction of symbolicdescriptions allows a much more sophisticated form of communication: structures are constructed from staticdescriptions and do not have to reproduce through some complicated, and limited process of self-inspection.In other words, separate descriptions can be used to reliably construct any kind of structure in an open-endedmanner, while self-inspection relies on only those structures that happen to be able to make copies ofthemselves. As an example, a non-genetic protein-based life form, would have to rely only on those proteinsthat could make direct copies of themselves [Rocha, 2001].

29

Further Readings and References:Belew, R.K. [1993]."Interposing an Ontogenic Model Between Genetic Algorithms and Neural Networks." In:

Advances in neural information processing (NIPS5). J. Cowan (Ed.). Morgan Kaufmann.De Jong, K.A.[2006]. Evolutionary Computation: A Unified Approach. MIT Press. Dellaert, F. and R.D. Beer [1994]."Toward an evolvable model of development for autonomous agent synthesis." In:

Artificial Life IV: Proceedings of the Fourth International Workshop on the Synthesis and Simulation of LivingSystems. R. Brooks and P. Maes (Eds.). MIT Press.

Goldberg, D. E. [1989]. Genetic Algorithms in Search, Optimization, and Machine Learning. Addison-Wesley.Gruau, Frédéric [1993]."Genetic Sythesis of Modular Neural Networks." In: Proceedings of the fifth international

conference on Genetic Algorithms. S. Forrest. Morgan Kauffman, pp 318-325.Hallinan, J. And J Wiles [2004]. “Asynchronous Dynamics of an Artificial Genetic Regulatory Network”. 9th

International Conference on Artificial Life. MIT Press. Holland, John H. [1975]. Adaptation in Natural and Artificial Systems. University of Michigan Press.C. Huang, J. Kaur, A. Maguitman, L.M. Rocha[2007].""Agent-Based Model of Genotype Editing". Evolutionary

Computation, 15(3).Kitano, H. [1990]."Designing Networks using Genetic Algorithms with Graph Generation System." In: Complex

Systems Vol. 4, pp 461-476.Kitano, Hiroaki [1994]."Evolution of Metabolism for Morphogenesis." In: Artificial Life IV: proceedings of the fourth

international workshop on the synthesis and simulation of living systems. R. Brooks and P. Maes (Eds.). MITPress.

Mitchell, M. [1992]. “Genetic algorithms”. In Lectures in Complex Systems. L. Nadel and D. Stein (Eds.). SFI Studiesin the Science of Complexity Vol. V, Addison-Wesley. pp 3-87.

Mitchell, M. [1999]. An Introduction to Genetic Algorithms. MIT Press.Reil, T. 1999. “Dynamics of Gene Expression in an Artificial Genome - Implications for Biological and Artificial

Ontogeny”. In Proceedings of the 5th European Conference on Advances in Artificial Life (September 13 - 17,1999). D. Floreano, J. Nicoud, and F. Mondada, Eds. Lecture Notes In Computer Science, vol. 1674.Springer-Verlag, London, 457-466

Rocha, Luis M. [1995]."Contextual Genetic Algorithms: Evolving Developmental Rules." In: Advances in ArtificialLife. F. Moran, A. Moreno, J.J. Merelo, and P. Chacon (Eds.). Series: Lecture Notes in Artificial Intelligence,Springer-Verlag. pp. 368-382.

Rocha, Luis M. [1996]."Eigenbehavior and symbols." In: Systems Research Vol. 12, No. 3, pp. 371-384.Rocha, Luis M. [1997]. Evidence Sets and Contextual Genetic Algorithms: Exploring Uncertainty, Context, and

Embodiment in Cognitive and Biological Systems. PhD. Dissertation. SUNY Binghamton.Rocha, Luis M. [1998]." Selected Self-Organization and the Semiotics of Evolutionary Systems." In: Evolutionary

Systems: The Biological and Epistemological Perspectives on Selection and Self- Organization. S. Salthe, G.Van de Vijver, and M. Delpos (eds.). Kluwer Academic Publishers, pp. 341-358.

Rocha, Luis M. [2001]. "Evolution with Material Symbol Systems". Biosystems. Vol. 60, pp. 95-121.Rocha, L.M., A. Maguitman, C. Huang, J. Kaur, and S. Narayanan. [2006].""An Evolutionary Model of Genotype

Editing". In: Artificial Life 10: Tenth International Conference on the Simulation and Synthesis of LivingSystems. L.M.Rocha, L. Yaeger, M. Bedau, D. Floreano, R. Goldstone, and A. Vespignani (Eds.). MIT Press,pp. 105-111.

Vasas, Vera, Eors Szathmary, and Mauro Santos [2010]. " Lack of evolvability in self-sustaining autocatalytic networks:A constraint on the metabolism-first path to the origin of life. ". Proceedings of the National Academy ofSciences of the United States of America: 0912628107.

Wagner, Gunter [1995] “Adaptation and the modular design of organisms”. In: Advances in Artificial Life. F. Moran,A. Moreno, J.J. Merelo, and P. Chacon (Eds.). Series: Lecture Notes in Artificial Intelligence, Springer-Verlag. pp. 317-328.

30